Ontological Engineering ©Asunción Gómez-Pérez,, M. Fernández, O. Corcho Ontological...

Ontological Engineering ©Asunción Gómez-Pérez,, M. Fernández, O. Corcho

Ontological Engineering:Methodologies and Tools

Asunción Gómez-PérezMariano Fernández-López

Oscar Corcho{asun, mfernandez, ocorcho}@fi.upm.es

Grupo de OntologíasLaboratorio de Inteligencia Artificial

Facultad de InformáticaUniversidad Politécnica de Madrid

Campus de Montegancedo sn,28660 Boadilla del Monte, Madrid, Spain


The Ontology Development ProcessMethodologies for building ontologiesMethods and tools for

ConceptualizingLearning ontologiesMergingEvaluatingEvolving

Outline


The Framework

The world of ontologies

• To set up a life cycle

• Development process

Tools

METHODOLOGYItem 1: It is necessary…

…….Item 2: Since …

Define-Ontology

(Imported ontologies ....)

ONTOLOGYCan be public

Gómez-Pérez, A. Knowledge Sharing and Reuse. In the Handbook of Applied Expert Systems. CRC Press. 1998.


Building ontologies

Import

Evaluate

Prune

Merge

Alignment

Identificar Diferencias

Specialize

Extend

Evolution

Export

Conceptualiz

e

Document

Integrate

Anotate

Reasoning

Specify

¿=?

+

O1O2

O3

...


Scheduling

Control

Qualityassurance

Management

Configurationmanagement

Knowledge acquisition

Evaluation

Documentation

Support

Integration

Specification Conceptualization

Formalization Implementation

Maintenance

Development oriented

Pre-development

Development

Post-development

Merging

Environment study Feasibility study

Use Alignment

Ontology Development Process


Specification Conceptualization Maintenance

Development activities

Support activities


Evaluation

Documentation

Configuration Management

Formalization

Integration

Implementation

Management activities

Scheduling Control

Quality assurance

Ontology Life Cycle


Inter-dependenciesInter-dependencies refer the relationship between activities carried out when building different ontologies

O1

O3

O2

Fernández-López, M.; Gómez-Pérez, A.; Rojas M.D.Ontology’s Crossed Life Cycle. Lectures Notes in Artificial Intelligence Nº 1937. October 2000


Methodologies and methods for building ontologies from scratch

Methods and Methodologies analysed (7):

• Cyc method

• Uschold and King’s method

• Grüninger and Fox’s methodology

• KACTUS method

• METHONTOLOGY

• SENSUS method

• On-To-Knowledge methodology

• Framework for comparing

methodologies

• Methodology/method description

• Comparison of the approaches

against the framework

• Conclusions


SENSUS as a basis for a domain-specific ontology (I)

Linking Domain Specific Terms to a broad Coverage Ontology

To identify the terms in SENSUS that are relevant to a particular domain and

then prune the skeletal ontology using heuristics

B. Swartout; R. Patil; k. Knight; T. Russ. Toward Distributed Use of Large-Scale OntologiesOntological Engineering. AAAI-97 Spring Symposium Series. 1997. 138-148.

SENSUS SENSUS

Skeletal Ontology


SENSUS as a basis for a domain-specific ontology (II)

Sensus Term

Seed

Path to root

Frequent Parent

Subtree Term

1. Identify “seed” terms

2. Link seed terms to SENSUS by hand

3. Include nodes on the path to root

4. Add entire subtrees using the heuristic:

If many nodes in a subtree are relevant,

the other nodes in the subtree are relevant

METHOD

B. Swartout; R. Patil; k. Knight; T. Russ. Toward Distributed Use of Large-Scale OntologiesOntological Engineering. AAAI-97 Spring Symposium Series. 1997. 138-148.


• Identify problem and opportunity areas

• Select most promising focus area and target solution

• Requirement specification

• Analyze input sources

• Develop baseline taxonomy

• Concept elicitation with domain experts

• Develop baseline taxonomy

• Conceptualize and formalize

• Add relations and axioms

• Identify problem and opportunity areas

• Select most promising focus area and target solution

• Manage organizationalmaintenance process

Project setting Ontology development

On-To-Knowledge


DocumentConfiguration Management

ControlQuality Assurance

Multilinguism

Prune

Extend

Specialize

RDF(S) DAML+OIL OWL

RDF(S) DAML+OIL OWL

Alignment

Merge

Evolution

Methontology

Specify

Conceptualize Evaluarte Implement

DAML+OIL

RDF(S)

OWL

Integrate Maintenance Use

Import

RDF(S) DAML+OIL OWL

Evaluate


Summary of the ontology development process

...


Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

To produce an Ontology Specification DocumentContent:

• Purpose• Scenarios of use• Possible end users• Level of formality of the ontology

• highly informal• semi-informal• semi-formal• rigorously formal

• Scope• Granularity

Language:• Informal • Semi-formal• Competency Questions


Getting terminology using Competency Questions

Identify intuitively possible applications and solutions

Identify Queries:• Answers: Axioms

Formal definitions• Questions: Terminology

MotivatingScenarios

InformalCompetency

Questions

Formal Terminology

ClassesRelationsAttributesAxiomsInstances

Uschold, M.; Grüninger, M. ONTOLOGIES: Principles, Methods and Applications. Knowledge Engineering Review. Vol. 11; N. 2; June 1996.

Find stories which include Person P

Identify Queries:• Questions: Story, Person, involved-in, includes • Answers: Story S1 includes person P

Classes: Story, PersonRelations: Involved-in, includesAttributes: ---AxiomsInstances: P, S1


Getting terminology using Competency QuestionsFind all the events attended by participants working on semantic web projects

Identify Queries:Questions: Event, Researcher, Project, work-at, attend-at, type-of-Project Answers: ISWC, EKAW, K-cap,....

Classes: Event, International Conference, National Conference, Workshop, Project Meetings, Researcher, Person, Project, Relations: Work-at, attend-atAttributes: Type of ProjectAxioms: For all...Instances: ISWC, EKAW, K-cap,....

Each project has a property storing its type

•Taxonomy of Topics•There exist a relation that connects projects and topics

Identify Queries:Questions: Event, Researcher, Project, work-at, attend-at, Semantic Web Topics, main-topics Answers: ISWC, EKAW, K-cap,....

Classes: Event, International Conference, National Conference, Workshop, Project Meetings, Researcher, Person, Project, Topics, Ontologies, mark-up languages, semantic web services, annotations, ...Relations: Work-at, attend-at, main-topics, topic-of Attributes: ---Axioms: For all ........ Instances:ISWC, EKAW, K-cap,....


Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

Ontology Development Process


METHONTOLOGY: Conceptualization

Gómez-Pérez, A. Knowledge Sharing and Reuse. In the Handbook of Applied Expert Systems. CRC Press. 1998.

It organizes and structures the knowledge acquired during the knowledge acquisition activity

using external representations that are independent of the knowledge representation

paradigms and implementation languages in which the ontology will be

formalized and implemented.

• We can use Ontology Editors for conceptualizing the Ontology

• The ontology editors transforms the conceptualization into executable code using translators


Task 1:Build glossary of terms

Task 2:Build concept taxonomies

Task 4:Build concept dictionary

Task 3:Build “ad-hoc” binary relation diagrams

Task 9:Describe formal axioms

Task 6:Describe instance

attributes

Task 7:Describe class

attributes

Task 8:Describeconstants

Task 5:Describe “ad-hoc”binary relations

Task 10:Describe rules

Task 11:Describe instances

Tasks of the conceptualization


Terms glossary


Primitives for Modelling Taxonomies

Disjoint decomposition: a set of subclasses of C that do not have common instances and do not cover C

Exhaustive-Decomposition: a set subclasses of C that cover C and may have common instances or subclasses

Partition: a set subclasses of C that cover C and do not have common instances or subclasses

Subclass-of:


Example of a Taxonomy (I)

Flight

Iberia FlightAmerican Airlines Flight British Airways Flight

AA7462 AA2010 AA0488 IB6274 BA0066 BA0069BA0068

Disjoint-DecompositionSubclass-ofSubclass-of

Subclass-of Subclass-of

Subclass-of

Subclass-of Subclass-ofSubclass-of


Example of a Taxonomy (II)

Flight

Domestic FlightInternational Flight

Partition

Economy Trip

Exhaustive-Decomposition

Travel Package

Business Trip Luxury Trip


Identify Ad-hoc relations

LocationTravel

arrival Place

is Arrival Place of

is Departure Place of

departure Place


Define a Concept Dictionary


Define in detail Instance Attributes


Define Class Attributes


Define formal axioms


Define rules


Define Instances


Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

Methods for reenginering ontologies

A possible Conceptual

Model

New Conceptual

ModelReestructuring:EvaluationRedesign

Configuration MangementReverse

EngineeringForward

Engineering

OntologyImplementation

New Ontology Implementation

Methods analysed (2):

•Method por reengineering integrated in METHONTOLOGY

•Onions proposes a method for reengineering ontologies


Ontology Libraries

DAML ontology library http://www.daml.org/ontologies/

Protege ontology library http://protege.stanford.edu/ontologies.html

Ontolingua ontology library http://ontolingua.stanford.edu/

WebOnto ontology library http://webonto.open.ac.uk

SHOE ontology library http://www.cs.umd.edu/projects/plus/SHOE/onts/index.html

WebODE ontology library http://webode.dia.fi.upm.es/

(KA)2 ontology library http://ka2portal.aifb.uni-karlsruhe.de/

AKT ontology http://www.aktors.org/ontology/


(def-class PUBLICATION-REFERENCE (abstract-information) "we have decided that a publication reference is an intangible, abstract information" ((has-title :type string) (has-author :type generic-agent) (has-date :type calendar-date) (has-place-of-publication :type location)))

(def-class ARTICLE-REFERENCE (Publication-Reference) ((has-page-numbers :type string) (article-of-journal :type journal) (issue-number :type integer) (issue-volume :type integer)))

(def-instance DKE-0169-023X (Article-Reference) (has-title “Methodologies, Tools and Languagesfor building ontologies: where is the meeting point?”) (has-author Corcho Fernández-López Gómez-Pérez) (has-date July-2003) (has-page-numbers 23) (article-of-journal DKE) (issue-volume 46))

Has-date

Generic-agentHas-author

Calendar-date

Has-place-of-publication

Location

JournalArticle-of journal

Publication-Reference

Article-Reference

Subclass-of

Subclass-of

Abstract-information

.- has-title: string

.- has-page-numbers: string

.- issue-number:integer

.- issue-volumen:integer

Has-title: “Methodologies, Tools and Languages for building ontologies: where is themeeting point?”) has-page-numbers: 23issue-volumen: 46

DKE

Instance-of

DKE-0169-023X

Instance-of

Corcho

Fernandez-Lopez

Gómez-Pérez

Instance-of

Instance-of

Instance-of

Article-of journal

Has-author

Has-author

Has-author


Selecting a tool for building the ontology

I must develop an ontology.

What Tool do I use to conceptualize it???

• The one(s) I like the most?

• The one(s) I know the best?

• The one(s) that import/export an ontology from/to a given ontology implementation language?

• The one(s) that best fit(s) my needs?


Main criteria for selecting an ontology editor Which activities of the ontology development process are supported by each tool?

What is the expressiveness of the underlying knowledge model attached to the tool?

What kinds of user interface does the tool provide to model ontology terms?

Does the tool provide an advanced user interface to model formal axioms or complex expressions?

Does the tool need to be installed locally or not?

Can it be used with a Web browser?

Where are the ontologies stored (in databases or files)?

Does the tool have an inference engine and querying tools?

Which ontology languages or formats does the tool generate?

Is the tool able to import ontologies implemented in ontology languages or in other formats?

Is it possible to export an ontology from one tool to another without losing knowledge?

How can ontology-based applications use ontologies developed with a tool?

What types of consistency checking and content evaluation does the tool perform?


Ontology development Tools

KAON from AIFB and FZI at the University of Karlsruhe http://kaon.semanticweb.org/

OilEd from University of Manchester http://oiled.man.ac.uk/

Ontolingua from KSL (Stanford University) http://www-ksl.stanford.edu

OntoSaurus from ISI (USA) http://www.isi.edu/isd/ontosaurus.html

OntoEdit from Karlsrhue Univ. http://ontoserver.aifb.unikarlsruhe.de/ontoedit/

Protégé 2000 from SMI (Stanford University) http://protege.stanford.edu/

WebOnto from KMI (Open University) http://kmi.open.ac.uk/projects/webonto/

WebODE from UPM http://webode.dia.fi.upm.es/webODE/


Ontology Development Tools

Ontology-Based Applications

Ontology Middleware

Ontology library

Ontologies

Metricsservices

Administrationservices

Ontology selectionservices

Queryservices

Ontology accessservices

...

Ontologyeditor

Ontologymerge

Ontologytranslation

SemanticPortals

KnowledgeManagement

Brokers ...Ontology

DevelopmentSuite

Component-basedEasy integrationRAD...

Ontologyacquisition

Ontologybrowser

Ontologyevaluation

Ontologyconf. man.

Ontologydocum.

Ontologyevolution alignment

Ontology



Example of Domain Ontology



Comparison of Ontology building tools

Criteria:

• General Description

• Tools’ architecture: architecture, extendibility, ontology storage, back-up

• Tools’ interoperability: with tools, export/import from/to languages

• KR paradigm supported by the tool

• Methodological Support

• Tools’ inference services

• Tools’ usability

• Framework for comparing tools

• Tool description

• Comparison of the tools


• Conclusions

• Recommendations


Ontology Development Tools. General description


Software architecture

Knowledge Representation Approach


Main Features of the editor and Inference Engine


Interoperability

Why low interoperability?

SIG3EON WS

Protégé-2000

RDF(S)

RDF(S) Ad hoc

WEbODEKnowledge Model

RDF(S)

Protégé-2000Knowledge Model RDF(S)

Se pierde


Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

Ontology Learning is the set of methods

and techniques used for building an ontology from scratch,

enriching, or adapting an existing ontology in

a semi-automatic fashion using several sources.

It aims to reduce the time and the effort necessary in

the knowledge acquisition process.

•Approaches:•Ontology learning from text

•Ontology learning from dictionary

•Ontology learning from knowledge bases

•Ontology learning from semi-structured schemata

•Ontology learning from relational schemata


Approaches for Ontology Learning

OL from text

• 18 methods

• 18 tools

OL from dictionary

• 3 methods

• 2 tools

OL from knowledge bases

• 1 method and tool

OL from semi-structured schemata

• 4 methods

• 1 tool

OL from relational schemata

• 4 methods

For each group of methods:

• Framework for comparing OL methods

• Method description

• Comparison of each Method


• Conclusions

• Recommendations

For each group of tools:

• Framework for comparing OL Tools

• Tool description

• Comparison of each Tool


• Conclusions

• Recommendations


Techniques used in different OL approaches

OL from text

• Natural Language Techniques

• Clustering techniques

• Machine learning

• Statistical aproach

OL from dictionary

• Natural Language Processing

• Statistical aproach


• Rules


• Graph Theory

• Machine Learning

• Pattern Recognition

• Clustering

• Ontological Techniques


• Mapping Techniques

• Reverse Engineering


OL from texts:methods and techniquesAguirre and colleagues’ method

Alfonseca and Manandhar’s method

Aussenac-Gilles and colleagues’ approach

Bachimont’s method

Faatz and Steinmetz approach

Gupta and colleagues’ approach

Hahn and colleagues’ method

Hearst’s approach

Hwang’s method

Khan and Luo’s method

Kietz and colleagues’ method

Lonsdale and colleagues’ method

Missikoff and colleagues’ method

Moldovan and Girju’s method

Nobécourt approach

Roux and colleagues’ approach

Wagner approach

Xu and colleagues’ approach

URL: Not available

URL: http://www.ii.uam.es/~ealfon

URL: http://www-lipn.univ-paris13.fr/~szulman/TERMINAE.html

URL: http://opales.ina.fr/public/

URL: Not available

URL: Not available

URL: Not available


URL: http://www.argreenhouse.com/InfoSleuth/index.shtml

URL: Not available

URL: http://ontoserver.aifb.uni-karlsruhe.de/texttoonto/

URL: http://www.ttt.org/salt/index.html

URL: Not available

URL: Not available

URL: Not available

URL: Not available

URL: Not available

URL: Not available


Hearst’s method

Rigau and colleagues’ method

Jannink and Wiederhold’s approach

URL: Not available

URL: http://www.lsi.upc.es/~rigau/

URL: Not available

OL from dictionary




Deitel and colleagues’ approach

Doan and colleagues approach

Papatheodorou and colleagues’ method

Volz and colleagues’ approach

URL: http://mondeca-publishing.com/s/anonymous/title11884.html

URL: Not available

URL: http://www.educanext.org/

URL: http://www.aifb.uni-karlsruhe.de/WBS/rvo/raphael-bib.html#

wonderweb-D11

Johannesson’s method

Kashyap’s method

Rubin and colleagues’ approach

Stojanovic and colleagues’ approach

URL: Not available

URL: Not available

URL: http://www.nigms.nih.gov/funding/pharmacogenetics.html

URL: http://wonderweb.semanticweb.org/publications.shtml


Criteria to descr ibe methods and techniques

• General Description, including its main goals and scope

• General steps used for learning

• Knowledge sources used for learning

• Main techniques applied in the process

• Possibility of reusing other ontologies

• Domains in which it has been tested

• Tools associated

• Most relevant ontologies built following it

• Bibliography

• URL


Comparison of OL methods from texts

...


Criteria followed to describe tools

• General Description including its main goals and scope

• Main techniques used by the tool

• Method followed

• Software architecture

• Interoperability with other tools

• Inport and export facilities

• Interface facilities

• URL

• Bibliography


OL from texts: tools18 tools described

ASIUM

Caméléon

Corporum-Ontobuilder

DOE

KEA

LTG

MO’K Workbench

OntoLearn

Prométhée

SOAT

SubWordNet E.P.

SVETLAN’

TDIDF

TERMINAE

TextStorm and Clouds

TextToOnto

Welkin

WOLFIE

URL: http://www.lri.fr/~faure/Demonstration/Presentation_Demo.html

URL: Not available

URL: http://ontoserver.cognit.no

URL: http://opales.ina.fr/public/

URL: http://www.nzdl.org/Kea/

URL: http://www.ltg.ed.ac.uk/%7Emikheev/workbench.html

URL: Not available

URL: Not available

URL: http://www.sciences.univ-nantes.fr/info/perso/permanents/morin/promethee/

URL: http://www.iis.sinica.edu.tw/IASL/en/index.htm

URL: http://www.aic.nrl.navy.mil/~aha/cbr/luikm.html

URL: http://www.limsi.fr/Individu/gael/ManuscritThese/

URL: Not available

URL: http://www-lipn.univ-paris13.fr/~szulman/TERMINAE.html

URL: Not available

URL: http://ontoserver.aifb.uni-karlsruhe.de/texttoonto/


URL: Not available


OL from texts. Tools

...


Conclusions about Ontology learning

• Ontology learning is a suitable process:

– to accelerate the knowledge acquisition process necessary to build an ontology from scratch,

– to reduce the time required to enrich an existing ontology,

– to speed up the construction of ontologies to be used for different purposes in the Semantic Web.

• integrated methods and techniques are needed for achieving the goal.


Ontology-based annotation tools

Ontology based annotation tools

• Used for Ontology population

• Main Features

• Language for storing the annotations

• Language for handling ontologies

• Automatization degree of the annotation process

• Static/dynamic page annotation

• Text/image annotation

AEroDAML

COHSE

MnM

OntoAnnotate

SHOE Knowledge Annotator



Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

Criteria:

Merging at run time or design time

Techniques used:

•Hierarchical clustering techniques

•FCA

•Terminological Analysis

Methods and Methodologies analysed (5):

•ONIONS,

•PROMPT,

•FCA-Merge,

•Information-Flow-based Ontology Mapping,

•The MOMIS methodology


FCA-Merge

Taxonomy 1

TAXONOMIES

Root

C1.1C1.2

C1.3C1.4

Root

C2.1

C2.2

C2.3Taxonomy 2

Doc. 1 Doc. 2 Doc. 1...DOCUMENTS

Root 1 C1.1 C1.2 C1.3 C1.4

Doc. 1

Doc. 2

...

Doc. n

X

X

X

X

X

X

X

X

X

X

X

X

X

Root 2 C2.1 C2.2 C2.3

Doc. 1

Doc. 2

...

Doc. n

X

X

X

X

X

X

X

X

X

X

X

X

CONTEXTS

( {doc.1,.., doc.3}, {C1.2, C2.1} )

({doc.1,.., doc. n}, {Root})

({}, {Bottom})

PRUNED LATTICE

R

MERGED ONTOL.

C1.2

Root


The Prompt MethodActivity 1.To make a list of suggested operations

Activity 5. To update the list of operations

Merge

Ontology O1 Ontology O2 Ontology O1 Ontology O2

Merge

Merge

Merge

Ontology O1 Ontology O2

Merge

Merge

Resulting ontology Resulting ontology

Activities 2 & 3. To select and toperform next operation

Conflict (e.g. data type missing)

Activity 4. To find conflicts !

It is supposed that copy is the operation proposed for the classes that will not be merged


Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

Ontology Evolution:The ability to manage ontology changes and their effects by creating and maintaining different variants of the ontology [Noy and Klein, 02].

Approaches1. METHONTOLOGY,

Activity during the life cycle [Fernández-López et al., 97] Identification of the elements to be controlled [Gómez-Pérez and Rojas, 99] Control of changes Generation of status reports.

2. Types of changes [Noy and Klein, 02].

3. Klein and Fensel [Klein and Fensel, 01]: Identification Change specification Transparent evolution

4. Stojanovic’s Process [Stojanovic et al., 02]:


Scheduling

Control

Qualityassurance

Management



Evaluation

Documentation

Support

Integration



Maintenance


Pre-development

Development

Post-development

Merging


Use Alignment

Criteria:

•Content Evaluation on taxonomies

•Criteria: consistency, completeness

Methods analysed (3):

•Gómez-Pérez approach for taxonomy evaluation

•OntoClean Method

•Ontological Constrains Manager (OCM)


Conclusions• There exist stable methodologies and tools for building ontologies, but

they do not cover all the process of the ontology development process.

– Methontology (the recommended methodology to ontology development by

FIPA )

– On-To-Knowledge

• There exist methods and tools for specific tasks

– Reengineering

– Collaborative construction

– Merging

– Evaluating

– Evolution

– Ontology Learning

• Integration of specific methods in methodologies are needed

• Technological support for the whole ontology development process


To know more about this topics

Ontoweb WP1: D1.1.1 http://www.ontoweb.org

WP1: D1.3 Survey on Tools

WP1: D1.4 Survey on methodologies

WP1: D1.5 Survey on ontology learning

OntoRoadMap

http://babage.dia.fi.upm.es/ontoweb/wp1/OntoRoadMap/index.html

Gómez-Pérez, A.; Fernández-López, M.; Corcho, O.

Ontological Engineering. Springer Verlag. 2003

Date post:	12-Jan-2016
Category:	Documents
Upload:	steven-heath
View:	215 times
Download:	0 times

Ontological Engineering ©Asunción Gómez-Pérez,, M. Fernández, O. Corcho Ontological...

Documents